Can an idea from MIT save US manufacturing?
There are tens of thousands of known compounds in the physical world. Finding the right combination to create a new material, say a super-tough or ultra-light metal, can take years, if not decades, of hit-and-miss experimentation.
Now, a project rooted at MIT aims to cut this time dramatically, and in doing so, accelerate the invention, development, and production of new goods and technologies. The effort is building a massive database of compounds and their properties, aiming to do for manufacturing what the Human Genome Project is doing for the biopharmaceutical industry — providing the data, tools, and understanding that will lead to breakthrough products.
The ultimate goal: reviving US manufacturing.
“If you look at our next generation of manufactured goods, you would be hard pressed to find more than a handful that are not material based,” said Cyrus Wadia, an assistant director with the White House Office of Science and Technology Policy. “We believe materials are going to fuel the economy of the future and solve some of society’s greatest challenges.”
The rapid development of new materials is considered vital to the future of US manufacturing, which increasingly relies on innovation and advanced products to stay ahead of foreign competitors. Simply put, new materials lead to new products, and the faster new materials can be commercialized, the faster new products can be brought to market.
To understand the importance of materials, think of the portable electronics used every day, such as cellphones. People are able to surf the Web on the subway because of materials invented in the 1970s that led to portable, long-lasting lithium-ion batteries that power cellphones, tablets, and laptops.
The Obama administration has made material sciences a key component of its strategy to boost manufacturing, folding the Massachusetts Institute of Technology project into its Materials Genome Initiative, launched in the end of 2011 with $63 million in funding for the first year. The White House initiative now includes at least 60 universities, research institutions, and companies, which are testing materials, creating new computational tools and algorithms, building the database of compounds, and training the next wave of materials scientists.
The MIT team and other university partners now work with the US Energy Department’s Lawrence Berkeley National Laboratory in Berkeley, Calif., which hosts supercomputers that perform complex calculations to create mathematical models that simulate how materials behave in nature. For example, at what point do they bend, break, or corrode?
What became today’s Materials Genome Initiative began in 2006 in the MIT lab of professor Gerbrand Ceder, a materials scientist who recognized that the exponential growth of computing power could be put to work to speed the development of new materials, much as it had been employed to map human DNA. For years, the development of new materials meant painstaking manual experiments and long hours of searching technical journals for experimental results and data — often jealously guarded by scientists doing the experiments.
This tedious process meant it could take decades to bring a new material to market. For example, Velcro, the fastener used in everything from shoes to garments to carrying cases, was patented in the 1950s, but wasn’t widely adopted for another 20 years. Teflon, the non-stick coating for cookware and other products, took even longer. Invented in the late 1930s, it wasn’t broadly commercialized until the early 1960s.
Ceder and his team aimed to cut this development time by at least half by creating a virtual laboratory where materials are mixed, matched, and tested at the speed of supercomputers. So far, the project has computerized over 30,000 known inorganic compounds, substances such as minerals and metals without elements, such as carbon, found in living organisms. Ceder predicts it will take at least five more years to develop a thorough database.
The task is time consuming because each compound exhibits a variety of properties, such as strength, density, energy, stability, and corrosion, which must be characterized in the database through computer simulations. And not all properties can be characterized electronically — yet. For example, scientists are struggling to develop models to describe how much pressure a material can take before its shape is permanently altered.
“I hope the project inspires people to create better models,” Ceder said. “Developing the method is hard and expensive, but the computing is easy once you get beyond that.”
Despite the challenges, the effort is already paying dividends. Vincent Chevrier, a product development engineer at 3M, uses the database weekly to determine which materials he wants to test for the St. Paul company’s lithium-ion battery projects. 3M creates lithium-ion batteries for a wide range of uses, but Chevrier declined to comment on the specific projects for which he has used the database.
Before the materials database became available online in 2011, Chevrier said he would spend days scouring journal articles to screen materials and often come up with less information than the database provides.
“Now you instantly have it in your browser,” he said. “It’s a huge increase in efficiency.”
About 3,000 people like Chevrier have registered to use the free database and more than 400 access it each month. Ceder said the target audience is research professionals employed by companies, many of which don’t have the money to afford this type of computing power on their own.
The open-source approach championed by Ceder and now the Materials Genome Initiative is perhaps just as revolutionary as the database itself. Scientific discoveries are sometimes shared in journals, which have costly subscriptions. Data is rarely included in the articles, so if researchers want to build on successful experiments, they must call and ask for the data. Often, Ceder says, the answer is no.
Companies typically protect their findings out of fear that competitors will snatch their ideas. Even Ceder admits there were fears early on among his Materials Project team that someone might log into their database, take the data, and claim it as their own.
Many scientists say a cultural shift from proprietary research to sharing data is necessary for the Materials Genome Initiative to succeed. Researchers benefit and companies save money by eliminating the repetition of cumbersome experiments.
“I’m a big believer that if you give people large amounts of data,” Ceder said, “they are going to do things with it you would have never predicted.”
Lithium-ion batteries were first invented in 1976 by Exxon Mobil scientist M.S. Whittingham.
The 1976 discovery didn’t lead directly to the rechargeble batteries used today, but it was the first step. Many improvements and variations have been made since then.
Now variations of this rechargeable battery are used in a wide range of everyday products, including Apple products. The batteries are also used in the Boeing 787 Dreamliner and the Chevy Volt.